ClawHub
Back to skill

Security audit

Twitter Video Download

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward Twitter/X video downloader whose network access, yt-dlp execution, and local file writes fit its stated purpose.

Install yt-dlp from a trusted source, download only media you intend to save, and choose the output folder carefully. Avoid putting usernames, passwords, or tokens in PROXY_URL because the script may print the full proxy value in logs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The release notes and usage description present the skill as a simple download utility but do not clearly warn users that it writes downloaded media to local storage and supports user-controlled output paths. In an agent context, that omission can mislead users about filesystem side effects, increasing the risk of unintended file writes or misuse in sensitive directories.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The skill includes broad natural-language trigger examples such as 'Download this Twitter video...' and says it will 'automatically call this skill,' which can increase the chance of unintended invocation by an agent in loosely matched conversations. In an agent ecosystem, over-broad phrasing can cause the skill to activate on user text that was descriptive or quoted rather than an explicit request, leading to unexpected network access and file writes.

Missing User Warnings

Low
Confidence
85% confidence
Finding
The top-level description says the skill downloads videos but does not clearly warn that media will be written to local storage at a user-specified or default location. In agent-driven contexts, omission of this side effect can reduce informed consent and surprise users with disk writes, especially if the skill is auto-invoked.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical
Code
suspicious.dangerous_exec
Location
scripts/download.mjs:7